JIB-09741; No of Pages 6 Journal of Inorganic Biochemistry xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Journal of Inorganic Biochemistry journal homepage: www.elsevier.com/locate/jinorgbio
Thermodynamics of binding of a sulfonamide inhibitor to metal-mutated carbonic anhydrase as studied by affinity capillary electrophoresis Yosuke Sato, Hitoshi Hoshino, Nobuhiko Iki ⁎ Graduate School of Environmental Studies, Tohoku University, 6-6-07 Aramaki-Aoba, Aoba-ku, Sendai 980-8579, Japan
a r t i c l e
i n f o
Article history: Received 19 February 2015 Received in revised form 13 June 2015 Accepted 14 June 2015 Available online xxxx Keywords: Carbonic anhydrase Sulfonamides Zn(II) Metal-substituted enzymes Affinity capillary electrophoresis
a b s t r a c t By affinity capillary electrophoresis (ACE), the thermodynamic binding constants of a sulfonamide (SA) inhibitor to bovine carbonic anhydrase II (CA) and metal mutated variants (M-CAs) were evaluated. 1-(4Aminosulfonylphenylazo)-2-naphthol-6,8-disulfonate was used as the SA in the electrophoretic buffer for ACE. The Scatchard analysis of the dependence of the electrophoretic mobility of native CA on the SA concentration provided the binding constant to be Kb = (2.29 ± 0.05) × 106 M−1 (at pH 8.4, 25 °C). On the other hand, apoCA showed far smaller value [Kb = (3.76 ± 0.14) × 102 M−1], suggesting that the coordination of SA to the ZnII center controlled the binding thermodynamics. The ACE of M-CAs showed the same behaviors as native CA but with different Kb values. For example, Co–CA adopting the same tetrahedral coordination geometry as native CA exhibited the largest Kb value [(2.55 ± 0.05) × 106 M−1] among the M-CAs. In contrast, Mn– and Ni–CA, which adopted the octahedral coordination geometry, had Kb values that were about two orders of magnitude lower. Because the hydrophobic cavity of CA around the active center pre-organized the orientation of SA, thereby fixing the ligating NH− moiety to the apex of the tetrahedron supported by three basal His3 of CA, metals such as Zn and Co at the center of M-CA gave the most stable CA–SA complex. However, pre-organization was not favored for octahedral geometry. Thus, pre-organization of SA was the key to facilitating the tetrahedral coordination geometry of the ZnII active center of CA. © 2015 Elsevier Inc. All rights reserved.
1. Introduction Metalloenzymes harness a specific metal ion to be accommodated in a peculiar coordination environment, facilitating their catalytic functions including hydrolysis, redox reactions, and isomerization [1–3]. No other metal species can effectively serve as the catalytic center. For example, carbonic anhydrase (CA) catalyzes the hydration of carbon dioxide and dehydration of bicarbonate through the ZnII center in a tetrahedral coordination geometry supported by three histidine residues (His3) of CA (Fig. 1a) [4]. Substitution of the ZnII center with other divalent metal ions, such as CdII, CuII, and HgII, leads to a substantial reduction or loss of catalytic function [5]. Thus, the metalloenzyme should harbor a specific metal species that is most appropriate for the coordination environment not only to facilitate the catalytic activity of the enzyme but also such that high kinetic and thermodynamic stability is retained. Substitution of the ligating groups around the active center of CA with other amino acid residues by mutation has shown that metal-binding selectivity arises from the coordination geometry and number to overwhelm the selectivity; this is called the Irving–Williams series, in which the thermodynamic stability of metal ions to small ligands intrinsically follows the order ⁎ Corresponding author. E-mail address: [email protected] (N. Iki).
MnII b FeII b CoII b NiII b CuII N ZnII [6]. Thus, the tetrahedral coordination environment is indispensable for CA to retain the kinetically labile ZnII as the catalytic center. Inhibitors that strongly inhibit enzymatic activity are commonly believed to be good analogs for the transition state of the catalytic reaction. If this were true, comparison of the binding ability of the inhibitor to a series of metallo variants of a metalloenzyme should also clarify the most suitable metal species to be accommodated by the coordination environment of the enzyme. In the case of CA, the highest-affinity class of inhibitors is the arylsulfonamides as shown by the small dissociation constants of the complex (in the picomolar to micromolar range) [4]. Also crystallographic analyses of the arylsulfonamide–CA complexes have clearly shown the resemblance to the transition state of CO2 hydrolysis and HCO− 3 dehydration (Fig. 1) [4,7,8]. Herein, we report the evaluation of the binding constant of SA to metallo variants of bovine CA II (M-CAs). To the best of our knowledge, quantitative evaluation of the binding constants for M-CAs is unprecedented. Our data revealed that metals that adopt tetrahedral coordination geometry, such as ZnII and CoII, formed thermodynamically more stable M-CA– SA complexes than MnII, NiII, CuII, and CdII. Moreover, the preorganization of SA through the hydrophobic cavity played an important role in setting up the NH− group of SA to occupy the apex of the tetrahedral coordination geometry of ZnII. Affinity capillary electrophoresis (ACE) [9–11] was used in this study to demonstrate the applicability
http://dx.doi.org/10.1016/j.jinorgbio.2015.06.011 0162-0134/© 2015 Elsevier Inc. All rights reserved.
Please cite this article as: Y. Sato, et al., Thermodynamics of binding of a sulfonamide inhibitor to metal-mutated carbonic anhydrase as studied by affinity capillary electrop..., J. Inorg. Biochem. (2015), http://dx.doi.org/10.1016/j.jinorgbio.2015.06.011
2
Y. Sato et al. / Journal of Inorganic Biochemistry xxx (2015) xxx–xxx
Fig. 1. Schematic drawings of the catalytic center of CA showing (a) transition state of the hydration/dehydration reaction and (b) binding of an arylsulfonamide.
of the method, including the ability to use CA samples without accurate determination of the concentration and the tolerance of the method to the presence of impurities such as isozymes as well as apo and native CAs in the M-CA sample due to incomplete substitution of ZnII.
Fig. 2) was prepared via a diazo-coupling reaction (see the Supplementary data). All other chemicals were of reagent grade. Deionized (d.i.) water prepared with an Elix Advantage 5 Water Purification System (Merck Millipore) was used throughout the study.
2. Experimental
2.3. ACE
2.1. Equipment
Prior to the ACE run, the capillary (total length L = 80.6 cm, effective length l = 71.9 cm) was thoroughly washed with a mixture of 0.01 M SDS, 0.01 M NaOH, d.i. water, and the electrophoretic buffer (25 mM Tris-phosphate buffer [pH 8.4] containing 0–7.62 mM SA1). The sample solution of CA, apoCA, or M-CA in 25 mM Tris–HCl (pH 8.4) was prepared by mixing the solutions of the components at room temperature for 1–12 h (see Section 1.8 of the Supplementary data for the exact conditions and compositions). It was injected into the anodic end of the capillary by applying 50 mbar for 3 s, and capillary electrophoresismediated separation was then implemented by applying 30 kV under absorption detection at 200 nm at the cathodic end. The observed electrophoretic mobility (μobs) was calculated by Eq. (1),
For measurement of absorption spectra, a Shimadzu UV-1800 UV– vis spectrophotometer was used. Capillary electrophoresis was carried out using an Agilent CE 7100 instrument equipped with a fused silica capillary (inner diameter = 50 μm; outer diameter = 375 μm) supplied by GL-Sciences Inc. A TOA-DKK pH meter (HM-25R) with a combined glass electrode (GST-5425C) was used to measure the pH. 1H NMR spectra were obtained using a Bruker DPX-400 spectrometer. A Yamato Coolnics Circulator CTE42W was used to maintain the temperature. 2.2. Materials The CA used throughout the study was from bovine serum (N 98% purity, cat. no. C3934) purchased from Sigma-Aldrich (St. Louis, MO, USA), the main content of which was identified as BCA II by capillary electrophoresis of the CA isozyme II from bovine erythrocytes (C2522) purchased from Sigma-Aldrich (see the Supplementary data). Metalsubstituted CAs were prepared by simply mixing a metal salt solution and apoCA obtained by dialysis of CA with dipicolinic acid [12] (see the Supplementary data for details). For dialysis of CA, a Visking seamless cellulose tubing (UC-24-32-100) was used. The sulfonamide inhibitor 1-(4-aminosulfonylphenylazo)-2-naphthol-6,8-disulfonate (SA1,
μ obs ¼ ðl LÞ=ðt m V Þ
ð1Þ
where tm is the migration time of a peak. The electro-osmotic mobility (μeo) was estimated using Eq. (1) for the electro-osmotic flow (EOF) peak. The observed electrophoretic mobility of CA (μ CA,obs) at a certain SA1 concentration in the electrophoretic buffer was calculated by subtracting μeo from μobs (Eq. (2)). μ CA;obs ¼ μ obs –μ eo
ð2Þ
Fig. 2. Structures of sulfonamides.
Please cite this article as: Y. Sato, et al., Thermodynamics of binding of a sulfonamide inhibitor to metal-mutated carbonic anhydrase as studied by affinity capillary electrop..., J. Inorg. Biochem. (2015), http://dx.doi.org/10.1016/j.jinorgbio.2015.06.011
Y. Sato et al. / Journal of Inorganic Biochemistry xxx (2015) xxx–xxx
3
Here, the mobility toward the cathode was defined as positive. The CE run was repeated approximately nine times by increasing the SA1 concentration stepwise until the μ CA value was saturated. The electrophoretic data were analyzed using Scatchard plots to obtain the thermodynamic binding constant of SA1 to CA (vide infra). The whole process was repeated three times.
3. Results and discussion 3.1. Design of the inhibitor Despite the broad structural diversity of the inhibitors examined to date, including small inorganic anions, alkyl sulfamates, and aryl SAs, the class of ligands with the highest affinity remains the aryl SAs [4]. Hence, we used this motif as the inhibitor in our study. In order to observe binding of the inhibitor to CA using ACE, the inhibitor must be large enough in terms of molecular size or the number of electric charges to alter the electrophoretic mobility of CA upon binding. To avoid causing synthetic and structural complexities by introducing macromolecules to aryl SAs, we added a large number of electric charges to SA. Using azocoupling, sulfanilamide and 2-naphthol-6,8-disulfonate were conveniently linked to afford the SA inhibitor 1-(4-aminosulfonylphenylazo)2-naphthol-6,8-disulfonate (SA1) having three negative charges from the two sulfonate groups and the one amide group. In the ACE of bovine CA II (BCA II), 4-aminosulfonylbenzoate (SA2, Fig. 2) was able to shift the electrophoretic peak, suggesting that a charge of − 2 was large enough to alter the mobility of the enzyme [11]. Taylor and others have reported the kinetics and equilibrium of complex formation between HCA C and SA3 (Fig. 2), a 5,7-disulfonate isomer of SA1; they found that the formation rate (k1) is 5.8 × 105 M−1 s−1, the dissociation rate (k−1) is 0.075 s−1, and the binding constant (Kb) is 6.9 × 106 M−1, demonstrating the rapid and stable binding equilibrium of SA3 to CA [13]. Because the binding energy of SA to CA is primarily dominated by coordination of the sulfonamide NH− to the ZnII center [14], and considering the structural integrity of CA among the origin, one may anticipate that the ability of SA1 to bind to BCA II in our case was similar to that of SA3. In fact, the Kb value of SA1 to BCA II was estimated to be 2.29 × 106 M−1 (vide infra).
Fig. 3. Typical electropherograms of ACE of native CA. Capillary: L = 80.6 cm, l = 71.9 cm, inner diameter = 50 μm, outer diameter = 375 μm; electrophoresis: V = 30 kV, i = 6–12 μA, λDetect = 200 nm; electrophoretic buffer: [SA1] = 0–158 μM, [Tris]T = 25 mM, pH 8.4 adjusted with H3PO4; sample: [CA] = 1.0 × 10−5 M (0.3 mg/mL), [Tris–HCl] = 25 mM, [MO] = 5.0 × 10−5 M, 5% acetone, pH = 8.37.
When the concentration of SA in the electrophoretic buffer increased from 0, the peak of CA shifted to the right, as exemplified by the case in which the concentration of SA1 was 7.9 μM; the EOF and MO peaks did not shift dramatically (Fig. 3b). This suggested that the SA in the buffer bound to and increased the negative charge of CA, thus enhancing the observed mobility μ CA,obs. Here, μ CA,obs was not a function of the concentration of CA in the sample zone but a function of the molar fraction (R) of the CA–SA complex in the total of CA species (Eq. (5)), given by Eq. (6).
3.2. ACE of native CA
R ¼ ½CA‐SA=½CAtotal
ð5Þ
The purpose of this experiment was to examine how the center metal species in M-CA affected the binding equilibrium of SA to CA (Eq. (3)) through evaluation of the binding constant Kb (Eq. (4)).
μ CA;obs ¼ ð1‐RÞμ CA þ Rμ CA–SA :
ð6Þ
CA þ SA⇄CA−SA
ð3Þ
Kb ¼ ½CA−SA=½CA½SA
ð4Þ
In this study, we used BCA II because of its availability. Hereinafter, CA denotes BCA II unless otherwise noted. Fig. 3a shows a typical electropherogram of a native CA sample at pH 8.4; this electropherogram consisted of three major peaks representing acetone as a marker of EOF, CA, and methyl orange (MO) as the internal standard for the injection volume. Minor peaks were also observed around the CA peak, possibly from the isozymes, which can be ignored as long as the mobility of the main CA peak is being evaluated. By subtracting the electro-osmotic mobility (μeo) from the observed electrophoretic mobility (μobs), the electrophoretic mobility of CA (μ CA) can be estimated to be −(6.65 ± 0.05) × 10−5 cm2 s−1 V−1. Here, mobility toward the cathode was defined as positive. Thus, the CA was electrophoretically driven toward the anodic side to show a net negative charge at pH 8.4, which was consistent with the isoelectric point (pI) value of 5.9 for CA [15].
Here μ CA–SA is the electrophoretic mobility of the CA–SA complex, which can be obtained only when all the CA species bind to SA, namely R = 1. In this particular situation, the observed mobility is saturated to be μCA,obs = μCA–SA, from which the μCA–SA value can be estimated. When the concentration of SA1 was 158 μM (Fig. 3c), the shift in the CA peak was saturated to give μCA–SA = −(9.43 ± 0.07) × 10−5 cm2 s−1 V−1. The R value could then be determined by the shift in the electrophoretic mobility (Δμobs), as follows. R ¼ Δμ obs =Δμ obs; max
ð7Þ
where Δμ obs ¼ μ CA;obs –μ CA
ð8Þ
Δμ obs; max ¼ μ CA–SA –μ CA :
ð9Þ
Combination of Eqs. (4) and (5) yielded the Scatchard equation (Eq. (10)). R=½SA ¼ K b ð1–RÞ:
ð10Þ
Please cite this article as: Y. Sato, et al., Thermodynamics of binding of a sulfonamide inhibitor to metal-mutated carbonic anhydrase as studied by affinity capillary electrop..., J. Inorg. Biochem. (2015), http://dx.doi.org/10.1016/j.jinorgbio.2015.06.011
4
Y. Sato et al. / Journal of Inorganic Biochemistry xxx (2015) xxx–xxx
Plotting R / [SA] against R yielded the Scatchard plot with a slope − Kb (Fig. 4, see also Fig. S3). Three independent ACE runs gave the Kb values, the average of which was Kb = (2.29 ± 0.05) × 106 M− 1 where the error indicates the standard deviation of the three Kb values. We will not discuss the absolute value of Kb because it is conditional (at pH 8.4) and dependent on the acid dissociation equilibria for the sulfonamide NH2 of SA (Eq. (11)) and the ZnII center of CA (Eq. (12)). H2 NSO2 Ar⇄Hþ þ − HNSO2 Ar
ðHis3 ÞZnðOH2 Þ⇄Hþ þ ðHis3 ÞZnðOH− Þ
ð11Þ
ð12Þ
where His3 denotes the three histidine residues of the CA, which support the ZnII center in the tetrahedral geometry. Unavailability of the reference value for this particular CA–SA pair is another reason that we cannot discuss the absolute value of Kb. Therefore, we will instead focus on the variations in Kb values among the M-CA enzymes. 3.3. Preparation of apoCA and the ACE Prior to the preparation of M-CAs, apoCA was prepared from native CA via dialysis using 2,6-dipicolinic acid to scavenge the center ZnII ion (see the Supplementary data). To confirm whether apoCA bound to SA1, ACE was implemented (Fig. S4). In the absence of SA1, the pattern of the electropherogram was almost the same as in the case of native CA, suggesting that removal of ZnII from the CAs (including the isozymes) did not significantly alter the electrophoretic mobility (μCA). In fact, the μ CA value was estimated to be −(6.48 ± 0.04) × 10− 5 cm2 s− 1 V− 1, which was only slightly smaller (in negative) than the μ CA value of native CA. This can be understood by assuming that the ZnII center in native CA has an OH− ligand and + 1 charge in total and that one of His3 in apoCA was protonated to have the same number of charges, + 1. A tiny fraction of the second His in apoCA may also be protonated, resulting in the slightly smaller absolute value of μCA. Addition of higher concentrations of SA1 to the electrophoretic buffer caused no appreciable shift in the apoCA peak, suggesting that the binding of SA1 to apo-CA was very weak. Increasing the concentration of SA1 up to 7.62 mM gradually shifted the peak, but
complete saturation of the μ CA,obs was not achieved (Fig. S4) to hinder the estimation of μCA–SA and then Scatchard analysis. Hence, the least square curve fitting to Eq. (13) derived from Eqs. (4) and (5) was applied to obtain a Kb of (3.76 ± 0.14) × 102 M−1, which was approximately four orders of magnitudes less than that of native CA. μ CA;obs ¼ ðμ CA þ K b ½SAμ CA–SA Þð1 þ K b ½SAÞ:
ð13Þ
This suggested that the majority of the binding energy of the enzyme–inhibitor complex could be attributed to the metal-SA coordination bond, consistent with the conclusion of a calorimetric study showing that the ZnII-N− bond should be the dominant interaction between CA and arylsulfonamides as compared to hydrogen bonding and hydrophobic interactions [14]. 3.4. ACE of M-CAs Prior to ACE, M-CAs were prepared by simply mixing the apoCA and metal ion at pH 8.4. First, we tested the case of Zn–CA reconstituted from apoCA and ZnII ion (Fig. S5), which showed the same ACE patterns as native CA; the higher the concentration of SA1 in the electrophoretic buffer, the higher (in negative) the μ CA,obs, implying the binding of SA to CA. In addition, the complete reconstitution of Zn–CA was confirmed by the absence of the apoCA peak in the region when SA1 was added up to a concentration of 158 μM. The Kb value of Zn–CA was (2.05 ± 0.05) × 106 M−1, which was fairly similar to the Kb value of native CA, suggesting the validity of the Scatchard analysis of the ACE patterns for the re-constituted M-CAs. ACE of other M-CAs (M = Mn, Ni, Co, Cu, or Cd) showed the retardation of the M-CA peak upon increasing the concentration of SA1 (Figs. S6–S10), suggesting the shift of the equilibrium (Eq. (3)) toward the formation of M-CA–SA. The Scatchard analysis yielded Kb values that were substantially higher than that of apoCA (Table 1), suggesting that the binding of SA was driven by the coordination of the sulfonamide NH− to the metal center. Moreover, the values were significantly different when the metal at the center was modified. Thus, we then examined how the metal species affected the Kb. Table 1 also shows the coordination geometry and coordination number (CN) of metal centers in metal variants for human CA II [16]. It is reasonable to assume that the CNs and geometries are applicable to those of the present bovine CA because of the high structural homology between human and bovine isoforms [4]. We found that metal centers with a high CN yielded lower Kb values. The Lewis acidity of this series of divalent metal ions should be strongly affected by the CN such that the Lewis acidity may seem to be the predominant factor. However, the observed reaction for Kb was in fact for the ligand exchange reaction between the pre-existing OH− (assuming complete dissociation in Eq. (12)) and the entering SA, as shown by Eq. (14), ðHis3 ÞMðOH2 Þn ðOH− Þ þ SA⇄ðHis3 ÞMðOH2 Þn ðSAÞ þ OH−
ð14Þ
Table 1 Binding constants (Kb) of M-CAs to SA, electrophoretic mobility μCA, and the coordination geometry and number (CN) at the metal centera. M-CA
Kb/M−1
−μCA/10−5 cm2V−1 s−1
Coordination geometryb
CNb
native MnII CoII NiII CuII ZnII CdII apo
(2.29 ± 0.05) × 106 (1.18 ± 0.15) × 104 (2.55 ± 0.05) × 106 (9.31 ± 0.40) × 103 (4.64 ± 0.39) × 104 (2.05 ± 0.04) × 106 (5.30 ± 0.08) × 105 (3.76 ± 0.14) × 102
6.65 ± 0.05 5.54 ± 0.21 6.64 ± 0.05 5.31 ± 0.10 6.57 ± 0.02 6.54 ± 0.08 5.91 ± 0.04 6.48 ± 0.04
Tetrahedral Octahedral Tetrahedral Octahedral Square pyramidal Tetrahedral –c
4 6 4 6 5 4 –c
a
Fig. 4. Typical Scatchard plot of native CA. Experimental conditions are given in the caption of Fig. 3.
b c
The Kb and μCA values were determined in 25 mM Tris-H3PO4, at pH 8.4, 25 °C. Reference [16]. Not available.
Please cite this article as: Y. Sato, et al., Thermodynamics of binding of a sulfonamide inhibitor to metal-mutated carbonic anhydrase as studied by affinity capillary electrop..., J. Inorg. Biochem. (2015), http://dx.doi.org/10.1016/j.jinorgbio.2015.06.011
Y. Sato et al. / Journal of Inorganic Biochemistry xxx (2015) xxx–xxx
where n is the number of coordinating water molecules and varied from 0 to 2 depending upon the CN of the metal. Based on the bond energy, the difference between M–O− in (His3)M(OH2)n(OH−) and M–N− in (His3)M(OH2)n(SA), namely the ΔH value for Eq. (14), should not differ among the M species, because a weak M–O− bond should yield a weak M–N−, and vice versa. Given that the CN of the metal center is conserved in Eq. (14), the ΔS should not vary among the metal species. Therefore, the significant variance in the Kb values among the various metal centers should be attributed to other factors, which would increase the stability of the products for CoII and ZnII and/or decrease the stability of the products for MnII and NiII. As stated earlier, strong ligands for the metalloenzyme represent the transition state analogs of the catalytic reaction. Therefore, M-CA–SA complexes with large Kb values should have a metal center that adopts the same coordination geometry as native CA. Hence, native CA and Zn–CA would be expected to have the largest Kb values due to the tetrahedral coordination geometry of ZnII. The largest value, in fact, was obtained with Co–CA, in which CoII adopts the tetrahedral coordination geometry as shown by X-ray crystallography [16]. Therefore, CoII seems to be the best metal center in terms of the catalytic activity. However, it has been reported that Co–CA exhibits ca. 50% of the catalytic activity of native CA, because coordination of bicarbonate expanded the coordination geometry to octahedral to reduce the rate of the product release [17]. Other M-CAs adopting other coordination geometries, such as octahedral and square pyramidal geometries, tended to have lower values, suggesting the primary importance of the coordination geometry. There are many crystallographic structures of CA–SA complexes available in the Protein Data Bank (PDB, see Table 4 in reference [4]). CA is known to possess primary and secondary hydrophobic pockets that can accommodate the phenylsulfonamide moiety and the rest of the SA molecule, respectively [18]. The space-filling model of the crystallographic structure of the CA–SA complex with p-benzenesulfonamide and a pendant amino acid (PDB ID 1CNW) [18] showed perfect fit of the ligand to the primary and secondary hydrophobic cavities (Fig. S11). This suggested that the benzenesulfonamide was preorganized to anchor the position of its NH− suitable for occupying the apex of the ZnII in the tetrahedral coordination geometry. In addition to the hydrophobic interactions, hydrogen bonding of the sulfonamide moiety and Thr199 of the CA skeleton (S = O･･･HN and NH−･･･O)
5
[14] has been reported to facilitate such pre-organization. Considering the structural integrity between BCA II and HCA II, the benzenesulfonamide moiety of SA1 would be expected to be pre-organized in the cavity of BCA II to facilitate the coordination of NH− to the tetrahedral ZnII center to yield the stable CA–SA complex (Fig. 5a). In the case of octahedral coordination geometry for Mn– and Ni–CAs, coordination sites other than His3 should be occupied with three small ligands, such as H2O and OH−, at the initial state (Eq. (14)). Because such ligands are too small to be bound by hydrophobic cavities, the position must be optimized so as to stabilize the octahedral geometry. The reaction with SA1 replaced the OH− ligand with the sulfonamide NH−, the position of which was not necessarily appropriate for the octahedral geometry because the pre-organization of SA1 by the cavities should anchor the NH− at the apex of the tetrahedral geometry. This should significantly distort the six-coordinate geometry to a partially crowded configuration of the ligands around the metal (Fig. 5b), leading to a weak bond between the metal and NH−. Thus, the preorganization of the SA is not advantageous for the octahedral coordination geometry to achieve lower Kb values. In the case of Cu–CA, the coordination geometry of CuII is a fivecoordinate square pyramid having one OH− and one H2O molecule on either side of the His3 (n = 1 in Eq. (14)) [16]. Although the position of the NH− is pre-organized so as to form the tetrahedral coordination geometry, the position of one remaining H2O ligand can be adjustable to make the most stable square pyramidal geometry by switching the positions of the four corners of the basal square and one apex. Additionally, the geometry may be switchable to a trigonal bipyramid [19]. These characteristics should result in the Kb value being higher than that of Mn– and Ni–CAs. In the case of Cd–CA, the Kb value was slightly lower than those of Co– and Zn–CAs. Assuming the tetrahedral coordination geometry supported by His3, the larger ionic radius of CdII (0.78 Å for CN 4) than that of ZnII (0.60 Å for CN 4) and CoII (0.58 Å for CN 4) [20] should make the position of the apex ligand OH− farther from the His3 plane. The shifted position may be somewhat incompatible with the position of NH− preorganized by the hydrophobic cavities, resulting in unstable binding. Thus, Cd–CA highlighted that, in addition to the coordination geometry, the ionic radius of the metal ion is also a factor determining the stability of the M-CA–SA complex. Thus, the position of the NH− moiety fixed by
Fig. 5. Schematic drawing of the active center of CA adopting (a) tetrahedral and (b) octahedral coordination geometries with the SA ligand pre-organized by the CA cavity. NHis denotes the imidazole nitrogen of histidine residue of CA.
Please cite this article as: Y. Sato, et al., Thermodynamics of binding of a sulfonamide inhibitor to metal-mutated carbonic anhydrase as studied by affinity capillary electrop..., J. Inorg. Biochem. (2015), http://dx.doi.org/10.1016/j.jinorgbio.2015.06.011
6
Y. Sato et al. / Journal of Inorganic Biochemistry xxx (2015) xxx–xxx
the pre-organization of the SA by the CA cavity would alter the Kb value depending on the coordination geometry of the center metal ion. In addition, the effects of the stereochemistry defined by the pre-organization may be more dramatic than discussed above. In Eq. (14), we assumed that the number of OH− moieties in the M-CAs was 1. The electrophoretic mobility of M-CA in the absence of SA1 (μ CA) is tabulated in Table 1. These data showed that the −μCA values of Mn–CA, Ni–CA, and Cd–CA were significantly lower than those of Zn–CA, Co–CA, and Cu–CA, indicating that the numbers of OH− ligands in Mn–, Ni–, and Cd–CAs were less than those of the other M-CAs. Therefore, the ligand substitution reaction with SA (Eq. (14)) for Mn, Ni, and Cd should be more favorable than that with Zn, Co, and Cu without considering the stereochemistry around the metal center. However, Mn– and Ni–CAs had significantly less binding affinity for SA than Zn– and Co–CAs, suggesting that the stereochemistry defined by the hydrophobic cavity overwhelmed the effects of the presence or absence of the coordinating OH− at the initial state. 4. Conclusion The present study successfully applied ACE to the evaluation of the Kb of SA1 to metal variants of CA to show the clear dependence on the center metal species; sharing the same tetrahedral coordination geometry, native CA, Zn–CA, and Co–CA had increased affinity for SA compared with other metal variants, which adopted an octahedral or square pyramidal coordination geometry at the center. These data strongly suggested that the pre-organization of the SA ligand by the hydrophobic pocket of CA facilitated the tetrahedral coordination geometry. On the other hand, pre-organization provided other M-CAs having CNs of five and six with an unfavorable stereochemistry, resulting in lower Kb values. Thus, ZnII and CoII were the most suitable metal cations for retention by the tetrahedral coordination geometry provided by His3 of CA and the SA ligand pre-organized by the CA cavity. We suggest that the His3 motif in CA evolved to harness the Lewis acidity of ZnII as the active center arising from the four-coordinate tetrahedral geometry. Abbreviations Ar aryl group ACE affinity capillary electrophoresis BCA bovine carbonic anhydrase
CA CE CN HCA SA SA1 SA2 SA3
carbonic anhydrase capillary electrophoresis coordination number human carbonic anhydrase sulfonamide 1-(4-aminosulfonylphenylazo)-2-naphthol-6,8-disulfonate 4-aminosulfonylbenzoate 1-(4-aminosulfonylphenylazo)-2-naphthol-5,7-disulfonate
Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jinorgbio.2015.06.011.
References [1] S.J. Lippard, J.M. Berg, Principles of Bioinorganic Chemistry, University Science Books, Mill Valley, California, 1994. [2] H.B. Kraatz, N. Metzler-Nolte, Concepts and Models in Bioinorganic Chemistry, Wiley, Weinheim, 2006. [3] W. Kaim, B. Schwederski, A. Klein, Bioinorganic Chemistry — Inorganic Elements in the Chemistry of Life: An Introduction and Guide, Wiley, Weinheim, 2013. [4] V.M. Krishnamurthy, G.K. Kaufman, A.R. Urbach, I. Gitlin, K.L. Gudiksen, D.B. Weibel, G.M. Whitesides, Chem. Rev. 108 (2008) 946–1051. [5] S. Lindskog, P.O. Nyman, Biochim. Biophys. Acta 85 (1964) 462–474. [6] K.A. McCall, C.A. Fierke, Biochemistry 43 (2004) 3979–3986. [7] P.A. Boriack-Sjodin, S. Zeitlin, D.W. Christianson, H.-H. Chen, L. Crenshaw, S. Gross, A. Dantanarayana, P. Delgado, J.A. May, T. Dean, Protein Sci. 7 (1998) 2483–2489. [8] K.M. Merz, R. Hoffmann, M.J.S. Dewar, J. Am. Chem. Soc. 111 (1989) 5636–5649. [9] L.Z. Avila, Y.H. Chu, E.C. Blossey, G.M. Whitesides, J. Med. Chem. 36 (1993) 126–133. [10] Y.H. Chu, L.Z. Avila, H.A. Biebuyck, G.M. Whitesides, J. Med. Chem. 35 (1992) 2915–2917. [11] I.J. Colton, J.D. Carbeck, J. Rao, G.M. Whitesides, Electrophoresis 19 (1998) 367–382. [12] J.B. Hunt, M.-J. Rhee, C.B. Storm, Anal. Biochem. 79 (1977) 614–617. [13] P.W. Taylor, R.W. King, A.S.V. Burgen, Biochemistry 9 (1970) 2638–2645. [14] V.M. Krishnamurthy, B.R. Bohall, C.-Y. Kim, D.T. Moustakas, D.W. Christianson, G.M. Whitesides, Chem. Asian. J. 2 (2007) 94–105. [15] M. Jonsson, E. Pettersson, Isoelectric fractionation, Acta Chem. Scand. 22 (1968) 712–713. [16] K. Håkansson, A. Wehnert, A. Liljas, Acta Crystallogr Sect. D 50 (1994) 93–100. [17] V. Yachandra, L. Powers, T.G. Spiro, J. Am. Chem. Soc. 105 (1983) 6596–6604. [18] P.A. Boriack, D.W. Christianson, J. Kingery-Wood, G.M. Whitesides, J. Med. Chem. 38 (1995) 2286–2291. [19] J.A. Hunt, M. Ahmed, C.A. Fierke, Biochemistry 38 (1999) 9054–9062. [20] R. Shannon, Acta Crystallogr Sect. A 32 (1976) 751–767.
Please cite this article as: Y. Sato, et al., Thermodynamics of binding of a sulfonamide inhibitor to metal-mutated carbonic anhydrase as studied by affinity capillary electrop..., J. Inorg. Biochem. (2015), http://dx.doi.org/10.1016/j.jinorgbio.2015.06.011
Contents lists available at ScienceDirect
Journal of Inorganic Biochemistry journal homepage: www.elsevier.com/locate/jinorgbio
Thermodynamics of binding of a sulfonamide inhibitor to metal-mutated carbonic anhydrase as studied by affinity capillary electrophoresis Yosuke Sato, Hitoshi Hoshino, Nobuhiko Iki ⁎ Graduate School of Environmental Studies, Tohoku University, 6-6-07 Aramaki-Aoba, Aoba-ku, Sendai 980-8579, Japan
a r t i c l e
i n f o
Article history: Received 19 February 2015 Received in revised form 13 June 2015 Accepted 14 June 2015 Available online xxxx Keywords: Carbonic anhydrase Sulfonamides Zn(II) Metal-substituted enzymes Affinity capillary electrophoresis
a b s t r a c t By affinity capillary electrophoresis (ACE), the thermodynamic binding constants of a sulfonamide (SA) inhibitor to bovine carbonic anhydrase II (CA) and metal mutated variants (M-CAs) were evaluated. 1-(4Aminosulfonylphenylazo)-2-naphthol-6,8-disulfonate was used as the SA in the electrophoretic buffer for ACE. The Scatchard analysis of the dependence of the electrophoretic mobility of native CA on the SA concentration provided the binding constant to be Kb = (2.29 ± 0.05) × 106 M−1 (at pH 8.4, 25 °C). On the other hand, apoCA showed far smaller value [Kb = (3.76 ± 0.14) × 102 M−1], suggesting that the coordination of SA to the ZnII center controlled the binding thermodynamics. The ACE of M-CAs showed the same behaviors as native CA but with different Kb values. For example, Co–CA adopting the same tetrahedral coordination geometry as native CA exhibited the largest Kb value [(2.55 ± 0.05) × 106 M−1] among the M-CAs. In contrast, Mn– and Ni–CA, which adopted the octahedral coordination geometry, had Kb values that were about two orders of magnitude lower. Because the hydrophobic cavity of CA around the active center pre-organized the orientation of SA, thereby fixing the ligating NH− moiety to the apex of the tetrahedron supported by three basal His3 of CA, metals such as Zn and Co at the center of M-CA gave the most stable CA–SA complex. However, pre-organization was not favored for octahedral geometry. Thus, pre-organization of SA was the key to facilitating the tetrahedral coordination geometry of the ZnII active center of CA. © 2015 Elsevier Inc. All rights reserved.
1. Introduction Metalloenzymes harness a specific metal ion to be accommodated in a peculiar coordination environment, facilitating their catalytic functions including hydrolysis, redox reactions, and isomerization [1–3]. No other metal species can effectively serve as the catalytic center. For example, carbonic anhydrase (CA) catalyzes the hydration of carbon dioxide and dehydration of bicarbonate through the ZnII center in a tetrahedral coordination geometry supported by three histidine residues (His3) of CA (Fig. 1a) [4]. Substitution of the ZnII center with other divalent metal ions, such as CdII, CuII, and HgII, leads to a substantial reduction or loss of catalytic function [5]. Thus, the metalloenzyme should harbor a specific metal species that is most appropriate for the coordination environment not only to facilitate the catalytic activity of the enzyme but also such that high kinetic and thermodynamic stability is retained. Substitution of the ligating groups around the active center of CA with other amino acid residues by mutation has shown that metal-binding selectivity arises from the coordination geometry and number to overwhelm the selectivity; this is called the Irving–Williams series, in which the thermodynamic stability of metal ions to small ligands intrinsically follows the order ⁎ Corresponding author. E-mail address: [email protected] (N. Iki).
MnII b FeII b CoII b NiII b CuII N ZnII [6]. Thus, the tetrahedral coordination environment is indispensable for CA to retain the kinetically labile ZnII as the catalytic center. Inhibitors that strongly inhibit enzymatic activity are commonly believed to be good analogs for the transition state of the catalytic reaction. If this were true, comparison of the binding ability of the inhibitor to a series of metallo variants of a metalloenzyme should also clarify the most suitable metal species to be accommodated by the coordination environment of the enzyme. In the case of CA, the highest-affinity class of inhibitors is the arylsulfonamides as shown by the small dissociation constants of the complex (in the picomolar to micromolar range) [4]. Also crystallographic analyses of the arylsulfonamide–CA complexes have clearly shown the resemblance to the transition state of CO2 hydrolysis and HCO− 3 dehydration (Fig. 1) [4,7,8]. Herein, we report the evaluation of the binding constant of SA to metallo variants of bovine CA II (M-CAs). To the best of our knowledge, quantitative evaluation of the binding constants for M-CAs is unprecedented. Our data revealed that metals that adopt tetrahedral coordination geometry, such as ZnII and CoII, formed thermodynamically more stable M-CA– SA complexes than MnII, NiII, CuII, and CdII. Moreover, the preorganization of SA through the hydrophobic cavity played an important role in setting up the NH− group of SA to occupy the apex of the tetrahedral coordination geometry of ZnII. Affinity capillary electrophoresis (ACE) [9–11] was used in this study to demonstrate the applicability
http://dx.doi.org/10.1016/j.jinorgbio.2015.06.011 0162-0134/© 2015 Elsevier Inc. All rights reserved.
Please cite this article as: Y. Sato, et al., Thermodynamics of binding of a sulfonamide inhibitor to metal-mutated carbonic anhydrase as studied by affinity capillary electrop..., J. Inorg. Biochem. (2015), http://dx.doi.org/10.1016/j.jinorgbio.2015.06.011
2
Y. Sato et al. / Journal of Inorganic Biochemistry xxx (2015) xxx–xxx
Fig. 1. Schematic drawings of the catalytic center of CA showing (a) transition state of the hydration/dehydration reaction and (b) binding of an arylsulfonamide.
of the method, including the ability to use CA samples without accurate determination of the concentration and the tolerance of the method to the presence of impurities such as isozymes as well as apo and native CAs in the M-CA sample due to incomplete substitution of ZnII.
Fig. 2) was prepared via a diazo-coupling reaction (see the Supplementary data). All other chemicals were of reagent grade. Deionized (d.i.) water prepared with an Elix Advantage 5 Water Purification System (Merck Millipore) was used throughout the study.
2. Experimental
2.3. ACE
2.1. Equipment
Prior to the ACE run, the capillary (total length L = 80.6 cm, effective length l = 71.9 cm) was thoroughly washed with a mixture of 0.01 M SDS, 0.01 M NaOH, d.i. water, and the electrophoretic buffer (25 mM Tris-phosphate buffer [pH 8.4] containing 0–7.62 mM SA1). The sample solution of CA, apoCA, or M-CA in 25 mM Tris–HCl (pH 8.4) was prepared by mixing the solutions of the components at room temperature for 1–12 h (see Section 1.8 of the Supplementary data for the exact conditions and compositions). It was injected into the anodic end of the capillary by applying 50 mbar for 3 s, and capillary electrophoresismediated separation was then implemented by applying 30 kV under absorption detection at 200 nm at the cathodic end. The observed electrophoretic mobility (μobs) was calculated by Eq. (1),
For measurement of absorption spectra, a Shimadzu UV-1800 UV– vis spectrophotometer was used. Capillary electrophoresis was carried out using an Agilent CE 7100 instrument equipped with a fused silica capillary (inner diameter = 50 μm; outer diameter = 375 μm) supplied by GL-Sciences Inc. A TOA-DKK pH meter (HM-25R) with a combined glass electrode (GST-5425C) was used to measure the pH. 1H NMR spectra were obtained using a Bruker DPX-400 spectrometer. A Yamato Coolnics Circulator CTE42W was used to maintain the temperature. 2.2. Materials The CA used throughout the study was from bovine serum (N 98% purity, cat. no. C3934) purchased from Sigma-Aldrich (St. Louis, MO, USA), the main content of which was identified as BCA II by capillary electrophoresis of the CA isozyme II from bovine erythrocytes (C2522) purchased from Sigma-Aldrich (see the Supplementary data). Metalsubstituted CAs were prepared by simply mixing a metal salt solution and apoCA obtained by dialysis of CA with dipicolinic acid [12] (see the Supplementary data for details). For dialysis of CA, a Visking seamless cellulose tubing (UC-24-32-100) was used. The sulfonamide inhibitor 1-(4-aminosulfonylphenylazo)-2-naphthol-6,8-disulfonate (SA1,
μ obs ¼ ðl LÞ=ðt m V Þ
ð1Þ
where tm is the migration time of a peak. The electro-osmotic mobility (μeo) was estimated using Eq. (1) for the electro-osmotic flow (EOF) peak. The observed electrophoretic mobility of CA (μ CA,obs) at a certain SA1 concentration in the electrophoretic buffer was calculated by subtracting μeo from μobs (Eq. (2)). μ CA;obs ¼ μ obs –μ eo
ð2Þ
Fig. 2. Structures of sulfonamides.
Please cite this article as: Y. Sato, et al., Thermodynamics of binding of a sulfonamide inhibitor to metal-mutated carbonic anhydrase as studied by affinity capillary electrop..., J. Inorg. Biochem. (2015), http://dx.doi.org/10.1016/j.jinorgbio.2015.06.011
Y. Sato et al. / Journal of Inorganic Biochemistry xxx (2015) xxx–xxx
3
Here, the mobility toward the cathode was defined as positive. The CE run was repeated approximately nine times by increasing the SA1 concentration stepwise until the μ CA value was saturated. The electrophoretic data were analyzed using Scatchard plots to obtain the thermodynamic binding constant of SA1 to CA (vide infra). The whole process was repeated three times.
3. Results and discussion 3.1. Design of the inhibitor Despite the broad structural diversity of the inhibitors examined to date, including small inorganic anions, alkyl sulfamates, and aryl SAs, the class of ligands with the highest affinity remains the aryl SAs [4]. Hence, we used this motif as the inhibitor in our study. In order to observe binding of the inhibitor to CA using ACE, the inhibitor must be large enough in terms of molecular size or the number of electric charges to alter the electrophoretic mobility of CA upon binding. To avoid causing synthetic and structural complexities by introducing macromolecules to aryl SAs, we added a large number of electric charges to SA. Using azocoupling, sulfanilamide and 2-naphthol-6,8-disulfonate were conveniently linked to afford the SA inhibitor 1-(4-aminosulfonylphenylazo)2-naphthol-6,8-disulfonate (SA1) having three negative charges from the two sulfonate groups and the one amide group. In the ACE of bovine CA II (BCA II), 4-aminosulfonylbenzoate (SA2, Fig. 2) was able to shift the electrophoretic peak, suggesting that a charge of − 2 was large enough to alter the mobility of the enzyme [11]. Taylor and others have reported the kinetics and equilibrium of complex formation between HCA C and SA3 (Fig. 2), a 5,7-disulfonate isomer of SA1; they found that the formation rate (k1) is 5.8 × 105 M−1 s−1, the dissociation rate (k−1) is 0.075 s−1, and the binding constant (Kb) is 6.9 × 106 M−1, demonstrating the rapid and stable binding equilibrium of SA3 to CA [13]. Because the binding energy of SA to CA is primarily dominated by coordination of the sulfonamide NH− to the ZnII center [14], and considering the structural integrity of CA among the origin, one may anticipate that the ability of SA1 to bind to BCA II in our case was similar to that of SA3. In fact, the Kb value of SA1 to BCA II was estimated to be 2.29 × 106 M−1 (vide infra).
Fig. 3. Typical electropherograms of ACE of native CA. Capillary: L = 80.6 cm, l = 71.9 cm, inner diameter = 50 μm, outer diameter = 375 μm; electrophoresis: V = 30 kV, i = 6–12 μA, λDetect = 200 nm; electrophoretic buffer: [SA1] = 0–158 μM, [Tris]T = 25 mM, pH 8.4 adjusted with H3PO4; sample: [CA] = 1.0 × 10−5 M (0.3 mg/mL), [Tris–HCl] = 25 mM, [MO] = 5.0 × 10−5 M, 5% acetone, pH = 8.37.
When the concentration of SA in the electrophoretic buffer increased from 0, the peak of CA shifted to the right, as exemplified by the case in which the concentration of SA1 was 7.9 μM; the EOF and MO peaks did not shift dramatically (Fig. 3b). This suggested that the SA in the buffer bound to and increased the negative charge of CA, thus enhancing the observed mobility μ CA,obs. Here, μ CA,obs was not a function of the concentration of CA in the sample zone but a function of the molar fraction (R) of the CA–SA complex in the total of CA species (Eq. (5)), given by Eq. (6).
3.2. ACE of native CA
R ¼ ½CA‐SA=½CAtotal
ð5Þ
The purpose of this experiment was to examine how the center metal species in M-CA affected the binding equilibrium of SA to CA (Eq. (3)) through evaluation of the binding constant Kb (Eq. (4)).
μ CA;obs ¼ ð1‐RÞμ CA þ Rμ CA–SA :
ð6Þ
CA þ SA⇄CA−SA
ð3Þ
Kb ¼ ½CA−SA=½CA½SA
ð4Þ
In this study, we used BCA II because of its availability. Hereinafter, CA denotes BCA II unless otherwise noted. Fig. 3a shows a typical electropherogram of a native CA sample at pH 8.4; this electropherogram consisted of three major peaks representing acetone as a marker of EOF, CA, and methyl orange (MO) as the internal standard for the injection volume. Minor peaks were also observed around the CA peak, possibly from the isozymes, which can be ignored as long as the mobility of the main CA peak is being evaluated. By subtracting the electro-osmotic mobility (μeo) from the observed electrophoretic mobility (μobs), the electrophoretic mobility of CA (μ CA) can be estimated to be −(6.65 ± 0.05) × 10−5 cm2 s−1 V−1. Here, mobility toward the cathode was defined as positive. Thus, the CA was electrophoretically driven toward the anodic side to show a net negative charge at pH 8.4, which was consistent with the isoelectric point (pI) value of 5.9 for CA [15].
Here μ CA–SA is the electrophoretic mobility of the CA–SA complex, which can be obtained only when all the CA species bind to SA, namely R = 1. In this particular situation, the observed mobility is saturated to be μCA,obs = μCA–SA, from which the μCA–SA value can be estimated. When the concentration of SA1 was 158 μM (Fig. 3c), the shift in the CA peak was saturated to give μCA–SA = −(9.43 ± 0.07) × 10−5 cm2 s−1 V−1. The R value could then be determined by the shift in the electrophoretic mobility (Δμobs), as follows. R ¼ Δμ obs =Δμ obs; max
ð7Þ
where Δμ obs ¼ μ CA;obs –μ CA
ð8Þ
Δμ obs; max ¼ μ CA–SA –μ CA :
ð9Þ
Combination of Eqs. (4) and (5) yielded the Scatchard equation (Eq. (10)). R=½SA ¼ K b ð1–RÞ:
ð10Þ
Please cite this article as: Y. Sato, et al., Thermodynamics of binding of a sulfonamide inhibitor to metal-mutated carbonic anhydrase as studied by affinity capillary electrop..., J. Inorg. Biochem. (2015), http://dx.doi.org/10.1016/j.jinorgbio.2015.06.011
4
Y. Sato et al. / Journal of Inorganic Biochemistry xxx (2015) xxx–xxx
Plotting R / [SA] against R yielded the Scatchard plot with a slope − Kb (Fig. 4, see also Fig. S3). Three independent ACE runs gave the Kb values, the average of which was Kb = (2.29 ± 0.05) × 106 M− 1 where the error indicates the standard deviation of the three Kb values. We will not discuss the absolute value of Kb because it is conditional (at pH 8.4) and dependent on the acid dissociation equilibria for the sulfonamide NH2 of SA (Eq. (11)) and the ZnII center of CA (Eq. (12)). H2 NSO2 Ar⇄Hþ þ − HNSO2 Ar
ðHis3 ÞZnðOH2 Þ⇄Hþ þ ðHis3 ÞZnðOH− Þ
ð11Þ
ð12Þ
where His3 denotes the three histidine residues of the CA, which support the ZnII center in the tetrahedral geometry. Unavailability of the reference value for this particular CA–SA pair is another reason that we cannot discuss the absolute value of Kb. Therefore, we will instead focus on the variations in Kb values among the M-CA enzymes. 3.3. Preparation of apoCA and the ACE Prior to the preparation of M-CAs, apoCA was prepared from native CA via dialysis using 2,6-dipicolinic acid to scavenge the center ZnII ion (see the Supplementary data). To confirm whether apoCA bound to SA1, ACE was implemented (Fig. S4). In the absence of SA1, the pattern of the electropherogram was almost the same as in the case of native CA, suggesting that removal of ZnII from the CAs (including the isozymes) did not significantly alter the electrophoretic mobility (μCA). In fact, the μ CA value was estimated to be −(6.48 ± 0.04) × 10− 5 cm2 s− 1 V− 1, which was only slightly smaller (in negative) than the μ CA value of native CA. This can be understood by assuming that the ZnII center in native CA has an OH− ligand and + 1 charge in total and that one of His3 in apoCA was protonated to have the same number of charges, + 1. A tiny fraction of the second His in apoCA may also be protonated, resulting in the slightly smaller absolute value of μCA. Addition of higher concentrations of SA1 to the electrophoretic buffer caused no appreciable shift in the apoCA peak, suggesting that the binding of SA1 to apo-CA was very weak. Increasing the concentration of SA1 up to 7.62 mM gradually shifted the peak, but
complete saturation of the μ CA,obs was not achieved (Fig. S4) to hinder the estimation of μCA–SA and then Scatchard analysis. Hence, the least square curve fitting to Eq. (13) derived from Eqs. (4) and (5) was applied to obtain a Kb of (3.76 ± 0.14) × 102 M−1, which was approximately four orders of magnitudes less than that of native CA. μ CA;obs ¼ ðμ CA þ K b ½SAμ CA–SA Þð1 þ K b ½SAÞ:
ð13Þ
This suggested that the majority of the binding energy of the enzyme–inhibitor complex could be attributed to the metal-SA coordination bond, consistent with the conclusion of a calorimetric study showing that the ZnII-N− bond should be the dominant interaction between CA and arylsulfonamides as compared to hydrogen bonding and hydrophobic interactions [14]. 3.4. ACE of M-CAs Prior to ACE, M-CAs were prepared by simply mixing the apoCA and metal ion at pH 8.4. First, we tested the case of Zn–CA reconstituted from apoCA and ZnII ion (Fig. S5), which showed the same ACE patterns as native CA; the higher the concentration of SA1 in the electrophoretic buffer, the higher (in negative) the μ CA,obs, implying the binding of SA to CA. In addition, the complete reconstitution of Zn–CA was confirmed by the absence of the apoCA peak in the region when SA1 was added up to a concentration of 158 μM. The Kb value of Zn–CA was (2.05 ± 0.05) × 106 M−1, which was fairly similar to the Kb value of native CA, suggesting the validity of the Scatchard analysis of the ACE patterns for the re-constituted M-CAs. ACE of other M-CAs (M = Mn, Ni, Co, Cu, or Cd) showed the retardation of the M-CA peak upon increasing the concentration of SA1 (Figs. S6–S10), suggesting the shift of the equilibrium (Eq. (3)) toward the formation of M-CA–SA. The Scatchard analysis yielded Kb values that were substantially higher than that of apoCA (Table 1), suggesting that the binding of SA was driven by the coordination of the sulfonamide NH− to the metal center. Moreover, the values were significantly different when the metal at the center was modified. Thus, we then examined how the metal species affected the Kb. Table 1 also shows the coordination geometry and coordination number (CN) of metal centers in metal variants for human CA II [16]. It is reasonable to assume that the CNs and geometries are applicable to those of the present bovine CA because of the high structural homology between human and bovine isoforms [4]. We found that metal centers with a high CN yielded lower Kb values. The Lewis acidity of this series of divalent metal ions should be strongly affected by the CN such that the Lewis acidity may seem to be the predominant factor. However, the observed reaction for Kb was in fact for the ligand exchange reaction between the pre-existing OH− (assuming complete dissociation in Eq. (12)) and the entering SA, as shown by Eq. (14), ðHis3 ÞMðOH2 Þn ðOH− Þ þ SA⇄ðHis3 ÞMðOH2 Þn ðSAÞ þ OH−
ð14Þ
Table 1 Binding constants (Kb) of M-CAs to SA, electrophoretic mobility μCA, and the coordination geometry and number (CN) at the metal centera. M-CA
Kb/M−1
−μCA/10−5 cm2V−1 s−1
Coordination geometryb
CNb
native MnII CoII NiII CuII ZnII CdII apo
(2.29 ± 0.05) × 106 (1.18 ± 0.15) × 104 (2.55 ± 0.05) × 106 (9.31 ± 0.40) × 103 (4.64 ± 0.39) × 104 (2.05 ± 0.04) × 106 (5.30 ± 0.08) × 105 (3.76 ± 0.14) × 102
6.65 ± 0.05 5.54 ± 0.21 6.64 ± 0.05 5.31 ± 0.10 6.57 ± 0.02 6.54 ± 0.08 5.91 ± 0.04 6.48 ± 0.04
Tetrahedral Octahedral Tetrahedral Octahedral Square pyramidal Tetrahedral –c
4 6 4 6 5 4 –c
a
Fig. 4. Typical Scatchard plot of native CA. Experimental conditions are given in the caption of Fig. 3.
b c
The Kb and μCA values were determined in 25 mM Tris-H3PO4, at pH 8.4, 25 °C. Reference [16]. Not available.
Please cite this article as: Y. Sato, et al., Thermodynamics of binding of a sulfonamide inhibitor to metal-mutated carbonic anhydrase as studied by affinity capillary electrop..., J. Inorg. Biochem. (2015), http://dx.doi.org/10.1016/j.jinorgbio.2015.06.011
Y. Sato et al. / Journal of Inorganic Biochemistry xxx (2015) xxx–xxx
where n is the number of coordinating water molecules and varied from 0 to 2 depending upon the CN of the metal. Based on the bond energy, the difference between M–O− in (His3)M(OH2)n(OH−) and M–N− in (His3)M(OH2)n(SA), namely the ΔH value for Eq. (14), should not differ among the M species, because a weak M–O− bond should yield a weak M–N−, and vice versa. Given that the CN of the metal center is conserved in Eq. (14), the ΔS should not vary among the metal species. Therefore, the significant variance in the Kb values among the various metal centers should be attributed to other factors, which would increase the stability of the products for CoII and ZnII and/or decrease the stability of the products for MnII and NiII. As stated earlier, strong ligands for the metalloenzyme represent the transition state analogs of the catalytic reaction. Therefore, M-CA–SA complexes with large Kb values should have a metal center that adopts the same coordination geometry as native CA. Hence, native CA and Zn–CA would be expected to have the largest Kb values due to the tetrahedral coordination geometry of ZnII. The largest value, in fact, was obtained with Co–CA, in which CoII adopts the tetrahedral coordination geometry as shown by X-ray crystallography [16]. Therefore, CoII seems to be the best metal center in terms of the catalytic activity. However, it has been reported that Co–CA exhibits ca. 50% of the catalytic activity of native CA, because coordination of bicarbonate expanded the coordination geometry to octahedral to reduce the rate of the product release [17]. Other M-CAs adopting other coordination geometries, such as octahedral and square pyramidal geometries, tended to have lower values, suggesting the primary importance of the coordination geometry. There are many crystallographic structures of CA–SA complexes available in the Protein Data Bank (PDB, see Table 4 in reference [4]). CA is known to possess primary and secondary hydrophobic pockets that can accommodate the phenylsulfonamide moiety and the rest of the SA molecule, respectively [18]. The space-filling model of the crystallographic structure of the CA–SA complex with p-benzenesulfonamide and a pendant amino acid (PDB ID 1CNW) [18] showed perfect fit of the ligand to the primary and secondary hydrophobic cavities (Fig. S11). This suggested that the benzenesulfonamide was preorganized to anchor the position of its NH− suitable for occupying the apex of the ZnII in the tetrahedral coordination geometry. In addition to the hydrophobic interactions, hydrogen bonding of the sulfonamide moiety and Thr199 of the CA skeleton (S = O･･･HN and NH−･･･O)
5
[14] has been reported to facilitate such pre-organization. Considering the structural integrity between BCA II and HCA II, the benzenesulfonamide moiety of SA1 would be expected to be pre-organized in the cavity of BCA II to facilitate the coordination of NH− to the tetrahedral ZnII center to yield the stable CA–SA complex (Fig. 5a). In the case of octahedral coordination geometry for Mn– and Ni–CAs, coordination sites other than His3 should be occupied with three small ligands, such as H2O and OH−, at the initial state (Eq. (14)). Because such ligands are too small to be bound by hydrophobic cavities, the position must be optimized so as to stabilize the octahedral geometry. The reaction with SA1 replaced the OH− ligand with the sulfonamide NH−, the position of which was not necessarily appropriate for the octahedral geometry because the pre-organization of SA1 by the cavities should anchor the NH− at the apex of the tetrahedral geometry. This should significantly distort the six-coordinate geometry to a partially crowded configuration of the ligands around the metal (Fig. 5b), leading to a weak bond between the metal and NH−. Thus, the preorganization of the SA is not advantageous for the octahedral coordination geometry to achieve lower Kb values. In the case of Cu–CA, the coordination geometry of CuII is a fivecoordinate square pyramid having one OH− and one H2O molecule on either side of the His3 (n = 1 in Eq. (14)) [16]. Although the position of the NH− is pre-organized so as to form the tetrahedral coordination geometry, the position of one remaining H2O ligand can be adjustable to make the most stable square pyramidal geometry by switching the positions of the four corners of the basal square and one apex. Additionally, the geometry may be switchable to a trigonal bipyramid [19]. These characteristics should result in the Kb value being higher than that of Mn– and Ni–CAs. In the case of Cd–CA, the Kb value was slightly lower than those of Co– and Zn–CAs. Assuming the tetrahedral coordination geometry supported by His3, the larger ionic radius of CdII (0.78 Å for CN 4) than that of ZnII (0.60 Å for CN 4) and CoII (0.58 Å for CN 4) [20] should make the position of the apex ligand OH− farther from the His3 plane. The shifted position may be somewhat incompatible with the position of NH− preorganized by the hydrophobic cavities, resulting in unstable binding. Thus, Cd–CA highlighted that, in addition to the coordination geometry, the ionic radius of the metal ion is also a factor determining the stability of the M-CA–SA complex. Thus, the position of the NH− moiety fixed by
Fig. 5. Schematic drawing of the active center of CA adopting (a) tetrahedral and (b) octahedral coordination geometries with the SA ligand pre-organized by the CA cavity. NHis denotes the imidazole nitrogen of histidine residue of CA.
Please cite this article as: Y. Sato, et al., Thermodynamics of binding of a sulfonamide inhibitor to metal-mutated carbonic anhydrase as studied by affinity capillary electrop..., J. Inorg. Biochem. (2015), http://dx.doi.org/10.1016/j.jinorgbio.2015.06.011
6
Y. Sato et al. / Journal of Inorganic Biochemistry xxx (2015) xxx–xxx
the pre-organization of the SA by the CA cavity would alter the Kb value depending on the coordination geometry of the center metal ion. In addition, the effects of the stereochemistry defined by the pre-organization may be more dramatic than discussed above. In Eq. (14), we assumed that the number of OH− moieties in the M-CAs was 1. The electrophoretic mobility of M-CA in the absence of SA1 (μ CA) is tabulated in Table 1. These data showed that the −μCA values of Mn–CA, Ni–CA, and Cd–CA were significantly lower than those of Zn–CA, Co–CA, and Cu–CA, indicating that the numbers of OH− ligands in Mn–, Ni–, and Cd–CAs were less than those of the other M-CAs. Therefore, the ligand substitution reaction with SA (Eq. (14)) for Mn, Ni, and Cd should be more favorable than that with Zn, Co, and Cu without considering the stereochemistry around the metal center. However, Mn– and Ni–CAs had significantly less binding affinity for SA than Zn– and Co–CAs, suggesting that the stereochemistry defined by the hydrophobic cavity overwhelmed the effects of the presence or absence of the coordinating OH− at the initial state. 4. Conclusion The present study successfully applied ACE to the evaluation of the Kb of SA1 to metal variants of CA to show the clear dependence on the center metal species; sharing the same tetrahedral coordination geometry, native CA, Zn–CA, and Co–CA had increased affinity for SA compared with other metal variants, which adopted an octahedral or square pyramidal coordination geometry at the center. These data strongly suggested that the pre-organization of the SA ligand by the hydrophobic pocket of CA facilitated the tetrahedral coordination geometry. On the other hand, pre-organization provided other M-CAs having CNs of five and six with an unfavorable stereochemistry, resulting in lower Kb values. Thus, ZnII and CoII were the most suitable metal cations for retention by the tetrahedral coordination geometry provided by His3 of CA and the SA ligand pre-organized by the CA cavity. We suggest that the His3 motif in CA evolved to harness the Lewis acidity of ZnII as the active center arising from the four-coordinate tetrahedral geometry. Abbreviations Ar aryl group ACE affinity capillary electrophoresis BCA bovine carbonic anhydrase
CA CE CN HCA SA SA1 SA2 SA3
carbonic anhydrase capillary electrophoresis coordination number human carbonic anhydrase sulfonamide 1-(4-aminosulfonylphenylazo)-2-naphthol-6,8-disulfonate 4-aminosulfonylbenzoate 1-(4-aminosulfonylphenylazo)-2-naphthol-5,7-disulfonate
Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jinorgbio.2015.06.011.
References [1] S.J. Lippard, J.M. Berg, Principles of Bioinorganic Chemistry, University Science Books, Mill Valley, California, 1994. [2] H.B. Kraatz, N. Metzler-Nolte, Concepts and Models in Bioinorganic Chemistry, Wiley, Weinheim, 2006. [3] W. Kaim, B. Schwederski, A. Klein, Bioinorganic Chemistry — Inorganic Elements in the Chemistry of Life: An Introduction and Guide, Wiley, Weinheim, 2013. [4] V.M. Krishnamurthy, G.K. Kaufman, A.R. Urbach, I. Gitlin, K.L. Gudiksen, D.B. Weibel, G.M. Whitesides, Chem. Rev. 108 (2008) 946–1051. [5] S. Lindskog, P.O. Nyman, Biochim. Biophys. Acta 85 (1964) 462–474. [6] K.A. McCall, C.A. Fierke, Biochemistry 43 (2004) 3979–3986. [7] P.A. Boriack-Sjodin, S. Zeitlin, D.W. Christianson, H.-H. Chen, L. Crenshaw, S. Gross, A. Dantanarayana, P. Delgado, J.A. May, T. Dean, Protein Sci. 7 (1998) 2483–2489. [8] K.M. Merz, R. Hoffmann, M.J.S. Dewar, J. Am. Chem. Soc. 111 (1989) 5636–5649. [9] L.Z. Avila, Y.H. Chu, E.C. Blossey, G.M. Whitesides, J. Med. Chem. 36 (1993) 126–133. [10] Y.H. Chu, L.Z. Avila, H.A. Biebuyck, G.M. Whitesides, J. Med. Chem. 35 (1992) 2915–2917. [11] I.J. Colton, J.D. Carbeck, J. Rao, G.M. Whitesides, Electrophoresis 19 (1998) 367–382. [12] J.B. Hunt, M.-J. Rhee, C.B. Storm, Anal. Biochem. 79 (1977) 614–617. [13] P.W. Taylor, R.W. King, A.S.V. Burgen, Biochemistry 9 (1970) 2638–2645. [14] V.M. Krishnamurthy, B.R. Bohall, C.-Y. Kim, D.T. Moustakas, D.W. Christianson, G.M. Whitesides, Chem. Asian. J. 2 (2007) 94–105. [15] M. Jonsson, E. Pettersson, Isoelectric fractionation, Acta Chem. Scand. 22 (1968) 712–713. [16] K. Håkansson, A. Wehnert, A. Liljas, Acta Crystallogr Sect. D 50 (1994) 93–100. [17] V. Yachandra, L. Powers, T.G. Spiro, J. Am. Chem. Soc. 105 (1983) 6596–6604. [18] P.A. Boriack, D.W. Christianson, J. Kingery-Wood, G.M. Whitesides, J. Med. Chem. 38 (1995) 2286–2291. [19] J.A. Hunt, M. Ahmed, C.A. Fierke, Biochemistry 38 (1999) 9054–9062. [20] R. Shannon, Acta Crystallogr Sect. A 32 (1976) 751–767.
Please cite this article as: Y. Sato, et al., Thermodynamics of binding of a sulfonamide inhibitor to metal-mutated carbonic anhydrase as studied by affinity capillary electrop..., J. Inorg. Biochem. (2015), http://dx.doi.org/10.1016/j.jinorgbio.2015.06.011
